interrogative -indefinites

[Koontz John E]

On Fri, 30 Sep 2005, shokooh Ingham wrote:
> One feature that I note is that (Dakota and Cree) both have, though not
> to the same extent, the phenomenon of interrogative-indefinites.
> Lakota has this to a very highly developed extent with its T-words taku
> 'what, something', tuwe 'who, someone', tuktel 'where, somewhere', tohan
> 'when, sometime' etc.  Cree has it but not to such a degree as Lakota
> and they even begin with T- in some cases.  I was wondering whether the
> interrogative-indefinites also ocur in other Siouan-Caddoan languages.
> ... So first question is do other Siouan languages have T-words.
> Second question is are they interrogatives, indefinites or
> interrogative-indefinites?

Hopefully the Caddoanists and Algonquianists on the list can clarify the
extent to which this pattern occurs elsewhere in those families.

In Mississipppi Valley Siouan I believe the pattern is universal.  I'd be
going out on a limb without looking to extend this to Crow-Hidatsa,
Mandan, and Southeastern.  However, the *ta- (and/or *to-) morpheme is not
always the base for interrogative/indefinites.  It does occur outside of
Dakotan, cf. Winnebago j^aagu' 'what' (j^aa- < *ta) or Omaha da'daN 'what'
(da- < *ta-), but other interrogative/indefinite bases occur.  (I think
*pe for 'who/someone' is pan-Siouan.)  The alternative to *ta-/*to- in
Dhegiha is *(h)a-, cf. OP a'naN 'how many/some number', or a'gudi
'where/somewhere'.  (The *h surfaces in Osage, Kaw, and Quapaw.  The same
rare initial pattern occurs with *(h)aNp- 'day', incidentally.)  The OP a-
forms are opposed to awa(N)- which is something like 'which of two'.  I
don't recall if this latter set have indefinite uses, but I'd expect them
to.  Not all a- forms match awaN- forms.  In fact, I think the lists are
somewhat skewed and essentially non-productive, though, since they often
include definite articles and postpositions, the lists are long.  I'm
thinking of forms like athedi/awaNthedi 'where/which place of two'.

Some question words look like they might be derived from standard
demonstrative forms, e.g., OP e?aN 'how' or eattaN - something like 'what
[unfortunate]'.  I think that these may involve focus constructions
historically, cf. French qu'est-ce que.  In other words, e?aN is 'how is
it that ...'.

There are some diffferences in interrogative/indefinite usage.  I think
that of the alternatives dadaN/edadaN/iNdadaN for 'what' the first is more
likely to be indefinite.  Also, I think there are some similar differences
for accent in e'be/ebe' 'who'.  I haven't noticed any similar patterns for
other forms, though there are some low frequency, speaker-restricted uses
of ?aN as 'how'.  (Perhaps a limited sort of dialect variation?)
Dorsey's texts are a nice place to look for differences in distribution.